As expressed by [[Bernhard Stadlbauer]] in his [[Untangling the Cloud - Choosing the Right Geospatial Format for Your Data Needs|talk]] for [[GeoPython 2025]]: > For a [[File Format|file format]] to be considered [[Cloud Computing|cloud]]-native, you should be able to perform partial reads. The common idea is that an index in the file lets you know what is available and exactly where in the file is the data you need.